How AI forces me to think better
Using AI has started to become my daily routine. I am still very careful with AI's reply as I am worried about hallucination. I never see AI as a threat but as a partner. Imagine AI as Jarvis from Ironman.
When I first use AI, I treat it like a chat bot. Then I started to copy AI prompt from the internet. Earlier this year, I started to let it perform data analysis. That's where I begin to have issues with AI's output.
- AI can be very convincing even though it is wrong.
- When your prompt is not clear, AI makes assumptions without telling you.
- AI never tells you that it does not know or do understand your instruction.
- Messy data means messy output. Sometimes it can get it right, sometimes it can get it wrong.
By using AI, it has forced me to think better:
- I need to be precise with what I want from AI.
- Even though AI can handle messy data structure, I still do my best to structure my data. A well-structured data system is easier to maintain and understand. The best ways is to learn the basics of database and data mining.
- I learned to challenge the results. The first output may not be correct, always challenge its finding.
- I learned to ask better questions. To ask better questions, I need to understand the topic better.
- Work with AI by telling you how it derives the answer that you asked.
If you can afford 2 subscriptions for AI agents, ask each agent to challenge the other agent's output. I have not try it, but I think it is a good idea.